1

I could PXE boot this laptop to some extent yesterday but all of a sudden after some time it started to show this error message. I reinstalled my tftp server etc in between.

It seems that DHCP lease issue is happening. The client wants some specific address I think? Does anyone know how to resolve this?

Client requested address 0.68.101.98 [06/06 04:37:27.658]
no more address or address previously allocated by another server [06/06 04:37:27.658]
Rcvd DHCP Discover Msg for IP 0.0.0.0, Mac 00:0E:7B:23:FA:11 [06/06 04:37:29.670]
Client requested address 0.68.101.98 [06/06 04:37:29.670]
no more address or address previously allocated by another server [06/06 04:37:29.670]
Rcvd DHCP Discover Msg for IP 0.0.0.0, Mac 00:0E:7B:23:FA:11 [06/06 04:37:33.679]
Client requested address 0.68.101.98 [06/06 04:37:33.679]
no more address or address previously allocated by another server [06/06 04:37:33.679]
Rcvd DHCP Discover Msg for IP 0.0.0.0, Mac 00:0E:7B:23:FA:11 [06/06 04:37:41.699]
1
  • 1
    What is doing DHCP? We need more info here.
    – DanBig
    Apr 8, 2010 at 19:29

2 Answers 2

1

This message usually occurs when your DHCP pool is full (i.e. every address that can be allocated by the DHCP server has been used, or is reserved).

If you restart the DHCP daemon (which should release all the IP addresses), does this still happen?

1
  • If resetting your dhcp daemon releases all IP addresses, that's a poor dhcp daemon. The ISC dhcpd (the one shipped with most linux distros these days) does not behave this way. Apr 9, 2010 at 4:02
0

That log doesn't quite make sense to me because it logs the Discover packet, but not the Offer and Request packets. And yet it says the client requested an address. But the address the client supposedly requested is clearly invalid. I don't know if the client is screwing up, or the DHCP server. It would be nice to see the reality of the packets on the wire.

Use a manageable switch to mirror the client laptop's port, put a sniffer on that port and look at all DHCP, ARP, and ICMP frames. You could use tcpdump something like this:

sudo tcpdump -i eth0 -nevvvs0 '(udp port bootpc) or arp or icmp'

Also, is this log really from June 6th? Have you been sitting on this question for 10 months, or is your system clock/calendar set wrong?

1
  • I second that. Check the clock in the BOIS of the PXE client. Oct 1, 2010 at 4:27

You must log in to answer this question.